How do you use / integrate org-roam in different setups

[leoc]

Hey there! I am new to the whole Zettelkasten/Roamresearch game, but I think I fell in love with the concept already! Thank you, for putting this thing out so I got wind of it!

Maybe you already have another channel for these discussions, but I would love to know how other people are integrating their existing notes, todos, project files etc with org-roam or similar. ❤️

I am currently reading the book from Sönke Ahrens and would love to incorporate org-roam into my workflow, but feel the need for some inspiration. 😅

[jethrokuan]

Maybe you already have another channel for these discussions, but I would love to know how other people are integrating their existing notes, todos, project files etc with org-roam or similar.

I don't yet, but if there's enough interest, I can start one.

[leoc]

Alright, so we can keep it here? :-)
I currently have a few giant "environment"-related (e.g. project, business, customer) org-files. I capture a lot of stuff into them. As far as I understand I would have to diverge from this paradigm and put everything into smaller files, so I can reference them? How do you manage your TODOs and do you integrate them into org-roam?

[jethrokuan]

put everything into smaller files, so I can reference them?

The basic idea is to split out your notes into standalone chunks, and tag them such that they surface when you want them to.

How do you manage your TODOs and do you integrate them into org-roam?

My TODOs are currently separate from my org-roam. I'm using just a few files, with org-agenda. (More here https://blog.jethro.dev/posts/org_mode_workflow_preview/). Org-agenda slows down massively with a large number of files.

[l3kn]

Making the org-agenda work with tasks / dates spread over a large number of files (> 1000) is tricky.

The main problem is that it uses markers for the locations of headlines,
these can't be persisted to disk and can't be created when parsing files in temp buffers
(which is faster than using the "real" files to do so).

A workaround (assuming only a small subset of files contains tasks / dates) is to set the org-agenda-files to a list of these files (e.g. using another cache or grep) before each call to org-agenda.

I think what I'll do for https://github.com/l3kn/org-zettelkasten (not working at the moment) is to make sure each task has an ID property and write wrappers for each org-agenda command that uses those instead of markers.

[jethrokuan]

I think what I'll do for https://github.com/l3kn/org-zettelkasten (not working at the moment) is to make sure each task has an ID property and write wrappers for each org-agenda command that uses those instead of markers.

oof, sounds rough.

A workaround (assuming only a small subset of files contains tasks / dates) is to set the org-agenda-files to a list of these files (e.g. using another cache or grep) before each call to org-agenda.

Yeah, I never noticed this, but I can always loop through files listed by the variable org-agenda-files for building the cache, in addition to what's in org-roam-directory. I don't think I'll ever find that useful though.

[CHaskell2]

Heya; I think this is the best place at the moment to ask a question like this, apologies if not.

I just started using org-roam and am looking into how to integrate it with pinboard for my daily capture/consume of bookmarks. I've got the trivial bit (use org-roam-today to create file, pull down bookmarks, append them as a plain text block to the note) working but I'm seeking some advice on a more complicated part.

So, pinboard has the concept of tagging your bookmarks; using that, this is what I'd ideally like to happen:

1. Add a link to pinboard with a tag
2. Bookmark is added to daily notes file
3. Bookmark references any topic files that match a tag from the bookmark
4. Topic files gain a link to the daily note

I'd set the processing of 2-4 on a daily cron, and I believe 4 would happen automatically if I do 3. right? For 3, I'd ideally avoid re-inventing org-roam-insert, but it also seems like the logic is straightforward (search for matching suffix "_topic.org", create org link to it, otherwise create file with timestamp and topic and link to that). Are there any beartraps I'd fall into if I reinvent the behavior of org-roam-insert in a separate process?

[nalhasan]

I, too, split org-agenda from org-roam because I don't want my writing to interfere with my planning/task management.

I suggest exploiting org-capture and refile tasks accordingly. You can have a header in one of your agenda files as "writing," "editing," or "revising" that links to the org-roam note you're currently in.

[alphapapa]

@CHaskell2 FYI you may find this useful for some of that: https://github.com/alphapapa/org-web-tools

[jethrokuan]

@CHaskell2 I don't really understand the workflow you're describing. I suppose you want to give your topic file tags a diiferent file name suffix?

[jethrokuan]

moving this kind of discussion to slack.

[aldanor]

A bit later to the party, but: https://d12frosted.io/posts/2021-01-16-task-management-with-roam-vol5.html

Basically, marking roam files with 'Project' tag, and then updating org-agenda-files to only look at those files, automatically. Isn't that cool? (any catches?..)

[d12frosted]

@Albanor I use this method with 3500 org files¹ and since most of them are not projects, org-agenda still takes less than a second to build (and I am also using custom sections and filters in it). The only catch, this approach is optimised towards many small files. If you have a 400k² lines file filled with TODOs, project marking is a little bit slow there. But since we are talking about org-roam, which shines with many small files approach, I guess it's not that critical. And it can be optimised further by not using elements-api (though this needs to be validated).

P. S. I am amazed someone reads that blog 😅

Footnotes

1. I finished the last bit of migration from org-brain to org-roam, that's why numbers are different :) ↩

2. Didn't perform proper benchmarks to find the size which makes things clumsy, I tested with 100k, 200k and 400k, each having 1/4 todo entries (e.g. 25k, 50k, 100k). ↩

[d12frosted]

@Albanor I forgot to mention few important things.

1. That approach depends on hooks working like a clock 😸 If they fail, some of the items might miss in your agenda. In my sense, this is terrible, because that's why we put all the tasks and thoughts into external system - our brain can't keep everything together. That's why two hooks and a script to go over all notes and make sure that project tag is up to date.
2. It uses advices. I know, advices are bad and stuff, but it does the work ideally. You could wrap org-agenda manually and use wrapper, but it has obvious drawback - not everyone is going to call your wrapper, meaning you might get into outdated agenda files list. Which is terrible, hence advice.

That being said, I am using this approach for more than a month and could not be more happy.

[Albanor]

Hi, @d12frosted. You misspelled the mention and summoned me instead of @aldanor ;)
That being said, I don't know if it's fate or not, I was interested a long time ago in note taking.
Thanks to your happy mistake I can now deep dive into the "Zettelkasten Method" (which I didn't know).
Cheers :D

[d12frosted]

🤦 While this turned out to be a happy mistake, I am sorry for misspelling your nicknames @Albanor and @aldanor (ah, even the first letter is of different case!). To my defense (and to retell a funny situation), I misspell my own name time to time, and once I did it thus far (email conversation):

• Me: I just noticed that I misspelled your name! I am terribly sorry!
• Bob: No worries, seriously. While we are at it, did you also misspell your name in email configurations?
• Me: 🤦 yep

So for a long time I was sending emails as Boiris instead of Boris

[zoranzaric]

@Albanor I forgot to mention few important things.

2. It uses advices.

What are advices in this context?

[d12frosted]

@zoranzaric Sorry, I was talking about Function Advising (I should've say advising instead of advice). Specifically, this:

(advice-add 'org-agenda :before #'+agenda-files-update)

So list of agenda files is updated before org-agenda is called.